The RAD52 Recombinational Repair Pathway is Essential in pol30 (PCNA) Mutants That Accumulate Small Single-Stranded DNA Fragments During DNA Synthesis

Corresponding author: Connie Holm, Department of Pharmacology, Center for Molecular Genetics, Division of Cellular and Molecular Medicine, University of California, San Diego, 9500 Gilman Drive, Mail Code 0651, La Jolla, CA 92093-0651, cholm{at}ucsd.edu (E-mail).

Communicating editor: F. WINSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To identify in vivo pathways that compensate for impaired proliferating cell nuclear antigen (PCNA or Pol30p in yeast) activity, we performed a synthetic lethal screen with the yeast pol30-104 mutation. We identified nine mutations that display synthetic lethality with pol30-104; three mutations affected the structural gene for the large subunit of replication factor C (rfc1), which loads PCNA onto DNA, and six mutations affected three members of the RAD52 epistasis group for DNA recombinational repair (rad50, rad52, and rad57 ). We also found that pol30-104 displayed synthetic lethality with mutations in other members of the RAD52 epistasis group (rad51 and rad54), but not with mutations in members of the RAD3 nor the RAD6 epistasis group. Analysis of nine different pol30 mutations shows that the requirement for the RAD52 pathway is correlated with a DNA replication defect but not with the relative DNA repair defect caused by pol30 mutations. In addition, mutants that require RAD52 for viability (pol30-100, pol30-104, rfc1-1 and rth1) accumulate small single-stranded DNA fragments during DNA replication in vivo. Taken together, these data suggest that the RAD52 pathway is required when there are defects in the maturation of Okazaki fragments.

THE proliferating cell nuclear antigen, PCNA or Pol30p in yeast, has been implicated in many of the pathways that allow rapid and precise replication of the genome. The crystal structure of PCNA reveals that it is a homotrimeric doughnut-shaped structure with a central pore suitable for surrounding double-stranded DNA molecules (KRISHNA et al. 1994 ). At high concentrations, PCNA is able to diffuse onto the ends of linear DNA molecules in vitro (BURGERS and YODER 1993 ), but to make this process efficient, replication factor C (RFC) must bind PCNA and place it around the DNA in an ATP-dependent reaction (FIEN and STILLMAN 1992 ; PODUST et al. 1995 ; TSURIMOTO and STILLMAN 1991 ). Once on the DNA, PCNA interacts with DNA polymerase or and enhances its ability to replicate DNA processively (BURGERS 1991 ; LEE and HURWITZ 1990 ; LEE et al. 1991 ; PODUST et al. 1992 ). In addition, PCNA physically interacts with FEN-1 (Rth1p/Rad27p/YKL510 in yeast), which is required for Okazaki fragment processing in vitro (ISHIMI et al. 1988 ; LI et al. 1995 ; TURCHI et al. 1994 ; WU et al. 1996 ). As expected from the biochemical activities ascribed to PCNA, DNA replication proceeds slowly in cold-sensitive pol30 mutants of Saccharomyces cerevisiae (AMIN and HOLM 1996 ). Somewhat surprisingly, the cold-sensitive pol30-104 mutant arrests at the G2/M stage of the cell division cycle with the bulk of its DNA replicated, but with an accumulation of single-stranded DNA gaps or nicks.

In addition to its role in DNA replication, PCNA is also important for several aspects of DNA repair. In vivo, mutations in POL30 cause sensitivities to genotoxic agents such as UV and MMS (methyl methane sulfonate) (AMIN and HOLM 1996 ; AYYAGARI et al. 1995 ). In vitro, PCNA is required for reconstitution of nucleotide excision repair and base excision repair reactions (MATSUMOTO et al. 1994 ; NICHOLS and SANCAR 1992 ; SHIVJI et al. 1992 ). Apparently, PCNA is required for the resynthesis of the approximately 30-bp ssDNA gap made during the excision of DNA damage. Epistasis analysis with the pol30-46 mutation suggests that this UV-sensitive and DNA-replication proficient mutant may be defective in the error-free post-replicative repair pathway controlled by Rad6p and Rad18p (TORRES-RAMOS et al. 1996 ).

Interestingly, PCNA also plays an important role in preventing the production of mutations. Two pol30 mutations, pol30-104 and pol30-52, cause mutator phenotypes that appear to be epistatic to mutations in members of the mismatch repair (MMR) pathway (JOHNSON et al. 1996 ; UMAR et al. 1996 ); these results suggest that PCNA may play a role in MMR. However, Pol30p may also play a role in other mutation avoidance pathways. Rth1p has also been placed in the MMR pathway because of its epistatic relationship with members of the MMR pathway (JOHNSON et al. 1996 ); further characterization of the mutator phenotypes caused by a rth1 mutation suggests that Rth1p may be important for a mutation avoidance pathway distinct from MMR (TISHKOFF et al. 1997 ). In addition, the synthetic lethality observed between rth1 and rad51/52 suggests that the RAD52 recombinational repair pathway may be required when there are defects in the RTH1 mutation-avoidance pathway (TISHKOFF et al. 1997 ). It is currently unknown whether PCNA also plays a role in the RTH1 mutation-avoidance pathway.

To further elucidate the role PCNA plays in these processes, we used a synthetic lethal screen to identify gene products required for viability when the functions of PCNA are impaired by the pol30-104 mutation. The isolation of rad50, rad52 and rad57 in this screen reveals that recombinational repair provides a compensatory mechanism by which pol30-104 mutants retain viability. The synthetic lethality observed in pol30 rad50 double mutants correlates with a disruption in the cell division cycle caused by the pol30 mutation. Moreover, mutations displaying synthetic lethality with rad52 cause an accumulation of small single-stranded DNA (ssDNA) fragments during DNA synthesis in vivo. Taken together, these observations suggest that the RAD52 pathway is important for viability when there are defects in Okazaki fragment maturation.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and plasmids:
All strains used in this study have an S288C background and are listed in Table 1. Standard genetic techniques and media were used in the construction and growth of each strain (SHERMAN et al. 1986 ). The integration of the pol30-104.LEU2 allele into strains CH1305 (MATa ade2 ade3 leu2 ura3-52 lys2-801) and CH1462 (MAT ade2 ade3 leu2 ura3-52 his3) to make strains CH2208 (MATa ade2 ade3 leu2 ura3-52 lys2-801 pol30-104.LEU2) and CH2210 (MAT ade2 ade3 leu2 ura3-52 his3 pol30-104.LEU2) was performed as described (AMIN and HOLM 1996 ). The plasmids used to construct rad51 (CH2307) and rad54 (CH2312) strains were kindly provided by DAVID SCHILD. Plasmids used to construct rad1 (CH2373), rad6 (CH2374) and rad14 (CH2375) strains were kindly provided by SATYA PRAKASH.

Cells were grown in YEPD (rich) or SD (minimal) medium. YEPD is 1% yeast extract, 2% bacto-peptone and 2% dextrose; SD is 0.67% yeast nitrogen base and 2% dextrose. For synthetic complete (SC) media, 20 mg of uracil, adenine, tryptophan and histidine, 30 mg of lysine, and 60 mg of leucine were added to 1 liter of SD media. 5-Fluoro-orotic acid (5-FOA) plates were made as described previously (AUSUBEL et al. 1988 ). All plates were made with 2% bacto-agar.

Plasmids were constructed using standard techniques of molecular biology (AUSUBEL et al. 1988 ) and are listed in Table 2. To make plasmid pCH1586 (POL30 HIS3 CEN/ARS) a 1.6-kb BamHI-XbaI fragment containing the entire POL30 gene plus approximately 600 bp of 5' untranslated sequence was ligated into the multi-cloning site (MCS) of BamHI-XbaI digested pCH1093 (HIS3 CEN/ARS) vector. We made use of other restriction endonuclease sites in the MCS of plasmid pCH1586 to construct the screening plasmid, pCH1589 (POL30 ADE3 URA3 CEN/ARS). A XhoI-NotI fragment containing the POL30 sequence of pCH1586 was ligated into Sal I-NotI digested plasmid pCH1153 (ADE3 URA3 CEN/ARS) to make plasmid pCH1589. Plasmid pCH1592 (POL30 HIS3 ADE3 CEN/ARS) was constructed by ligating a 3.8-kb BamHI-XhoI fragment containing the ADE3 gene from pCH1023 into the BamHI-Sal I sites of pCH1586.

Synthetic lethal screen:
The synthetic lethal screen used in this work is based on the plasmid dependence assay developed by KRANZ and HOLM 1990 . To perform the screen, strain CH2208 (MATa ade2 ade3 ura3 leu2 lys2 pol30-104.LEU2) or CH2210 (MAT ade2 ade3 ura3 leu2 his3 pol30-104.LEU2) was transformed with the screening plasmid pCH1589 (POL30 ADE3 URA3 CEN/ARS). Transformants were grown at 35° in selective medium to mid-log phase and plated onto YEPD plates. The plates were placed in a UV Stratalinker 1800 (Stratagene, La Jolla, CA) and mutagenized to ~80% viability (80–100 J/m²). Approximately 200,000 mutagenized colonies were assayed for sectoring after growing for 7 days at 35°. Non-sectored red colonies were struck for single colonies onto fresh YEPD and 5-FOA plates to confirm the dependence on pCH1589 for viability.

Secondary and tertiary tests were performed to identify non-sectoring mutants that had single-gene mutations causing a dependency on POL30 for viability. To determine if non-sectoring mutants were dependent on the POL30 gene on pCH1589 for viability, mutants were transformed with pCH1093 (HIS3) or pCH1586 (POL30 HIS3). Mutants that regained the ability to sector on YEPD and grow on 5-FOA when transformed with pCH1586 (POL30 HIS3) but not when transformed with pCH1093 (HIS3) were determined to be dependent on POL30 for viability. Mutants were subsequently analyzed to determine if a mutation causing POL30 dependence was located at a single locus by backcrossing with unmutagenized strain CH2208 (MATa ade2 ade3 ura3 leu2 lys2 pol30-104.LEU2) or strain CH2210 (MAT ade2 ade3 ura3 leu2 his3 pol30-104.LEU2).

Phenotypic characterization of POL30-dependant mutants:
To determine if any of the mutations causing inviability in a pol30-104 background also caused defects in a POL30/pol30 background, non-sectoring strains carrying plasmid pCH1589 (POL30 ADE3 URA3 CEN/ARS) were assayed for sensitivities to temperature and DNA damage. POL30 dependent mutants were spotted onto YEPD plates placed at 37°, 30°, 25°, 20° and 14° and on YEPD plates containing 0.02% MMS or 0.04 M HU at 30°.

Cloning of slp (s ynthetic lethal with pol30) mutations by suppression:
Since the slp2 and slp3 mutations caused sensitivities to HU and MMS, we identified these genes by isolating genomic library plasmids that suppressed the HU and MMS sensitivities. An URA3 CEN/ARS genomic library (pCH1020) (ROSE et al. 1987 ) was transformed into strains CH2286 (slp2-1) and CH2297 (slp3). Transformation plates were incubated at 35° for 2 days then replica-plated onto YEPD plates containing either 0.04 M HU or 0.02% MMS and incubated an additional 2 days at 35°. Plasmid DNA from HU and MMS resistant colonies was isolated as described (ROBZYK and KASSIR 1992 ) and retransformed into strain CH2286 (slp2-1) or CH2297 (slp3). Library plasmids that suppressed both the POL30 dependence and the HU and MMS sensitivities of CH2286 (slp2-1) or CH2297 (slp3) were partially sequenced by the dideoxy chain-termination method (Sequenase reagent kit from United States Biochemical Corp., Cleveland, OH). Approximately 100 bp of DNA sequence from each clone was submitted to the Saccharomyces Genome Database (http://genome-www. stanford.edu) for BLAST comparison against the yeast genome. Conventional subcloning techniques were used to generate nested deletions in each insert to identify regions of the library inserts required for suppression.

Specificity of observed synthetic lethality:
To determine if pol30-104 displayed synthetic lethality with different rad mutations, strain CH2161 (pol30-104.LEU2) was crossed with strain CH1637 (rad52::URA3), and strain CH2321 (pol30-104.TRP1) was crossed with strains CH2307 (rad51::LEU2) and CH2312 (rad54LEU2). Strain CH2319 (pol30-104.TRP1) was crossed with strains CH2373 (rad1::LEU2), CH2374 (rad6::hisG-URA3-hisG) and CH2375 (rad14::URA3). To determine if rad52 displayed synthetic lethality with DNA replication/repair genes other than pol30-104, strain CH1637 (rad52) was crossed with strains CH2359 (rfc1-5) CH2145 (pol1-16), CH2152 (cdc17-1), CH526 (cdc2-1), CH2234 (pol2-11) and CH2235 (pol2-12). Diploids were selected, sporulated, and germinated at 25° for pol1, pol2, cdc2 and cdc17 mutants, or 35° for pol30-104 and rfc1-5 mutants. The number of expected double mutants was determined by examining the phenotypes of all tetrads in which two or more spores survived.

pol30 allele-specific synthetic lethality:
Strain CH2335 {(pol30:: LEU2 rad50::hisG leu2 trp163 ura3-52 [pCH1511 (URA3 POL30)]} was transformed with TRP1 plasmids containing different pol30 alleles (Table 2). Transformants were grown in synthetic medium to select for both the POL30 URA3 and the pol30-x TRP1 plasmids. Stationary cultures and serial dilutions were spotted onto SD plates (to determine relative concentration of cells spotted) and 5FOA plates [to evaluate growth without the pCH1511 (POL30 URA3) covering plasmid].

Repair capacity of pol30 mutants:
To compare the relative sensitivities of the pol30 mutants to DNA damage, mutants were assayed for growth on either YEPD plates containing different amounts of MMS or YEPD plates exposed to UV in a Stratalinker 1800. Stationary cultures of pol30::LEU2 strains containing the indicated pol30 alleles on CEN/ARS plasmids were serially diluted and spotted onto YEPD plates. To determine the sensitivities of pol30 mutants to UV, these plates were irradiated with 0, 30, 55, 80, 100 or 150 J/m² of UV light. To determine the sensitivities of pol30 mutants to MMS, pol30 mutants were also spotted onto YEPD plates containing 0.005%, 0.01%, 0.02% or 0.03% MMS. Plates were incubated at 25° or 35° for 2 days.

Flow cytometry:
Cultures of pol30::LEU2 strains containing the indicated pol30 alleles on CEN/ARS plasmids (Table 2) were grown to log phase at 35°. An aliquot of the culture was taken and fixed in 70% ethanol. The remaining 35° culture was split and incubated at 25° or 14°. Aliquots were taken from the 25° culture after 4.5 hr and from the 14° culture after 18 hr, and then fixed in 70% ethanol. Fixed cells were sonicated, treated with 1 mg/ml Rnase A, and stained with 50 µg/ml propidium iodide prior to microfluorometric analysis (FACS) on a Becton Dickinson fluorescence activated cell sorter.

Alkaline sucrose velocity sedimentation gradients:
To determine if mutations affect the processing of Okazaki fragments, DNA from pulse-labeled cultures was resolved by alkaline sucrose velocity sedimentation. To focus our analysis solely on chromosomal DNA replication, all these experiments were performed with rho° strains lacking mitochondrial DNA (SHERMAN et al. 1986 ). The protocol used for these experiments was derived from a combination of previously published methods ( JOHNSTON and WILLIAMSON 1978 ; MCALEAR et al. 1996 ). Total DNA was chronically labeled by inoculating approximately 10⁴ cells in 1 ml of YEPD medium containing 0.08 µCi ¹⁴C-uracil, and incubating this culture overnight at permissive temperature. The log-phase cultures were transferred to 1.5 ml tubes and concentrated to 0.5 ml. Prior to addition of the pulse-label (100 µCi ³H-uracil), cultures were incubated for 10 min at 35° or 37°, or 20 min at 14°. Pulse-labeling times were 30 min at 35° or 37°, or 3 hr at 14°. Pulse-labeling was terminated by adding 0.6 ml of stop mix (3% toluene, 95% ethanol, 20 mM Tris-HCl pH 7.4, 20 mM EDTA) ( JOHNSTON and WILLIAMSON 1978 ).

DNA was isolated from the fixed cells as previously described (MCALEAR et al. 1996 ). After washing twice with ice cold buffer 1 (0.1 M Tris pH 8.5, 0.1 M EDTA, 2% 2-mercaptoethanol), cells were resuspended in 150 µl of cold buffer 2 (0.01 M KPO4 pH 9.72, 0.01 M EDTA) and transferred to a 5 ml pollyallomer tube. Zymolyase (20 µl of 10% solution) and Rnase A (10 µl of 10 mg/ml solution) were added to the samples, and they were incubated at 37° for 15 min. Twenty microliters each of NP-40 (10% solution) and sarkosyl (20% solution) and 25 µl of 5 M NaOH were added to the samples. Lysed cells were then floated on top of a 5%-20% sucrose gradient containing 0.7 M NaCl, 0.03 M EDTA and 0.3 M NaOH. Gradients were centrifuged in a Beckman SW55 rotor at 15,000 rpm for 17 hr. After centrifugation, twenty 250 µl fractions were collected, and the acid precipitable counts for each fraction were determined as previously described (MCALEAR et al. 1996 ). Gradient fraction #1 represents the top the gradient and fraction #20 is the bottom of the gradient.

The size of ssDNA in each gradient fraction was determined by using the equation (LEVIN and HUTCHINSON 1973 )

JOHNSTON and WILLIAMSON 1978

JOHNSTON and NASMYTH 1978

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Identification of mutations causing synthetic lethality with pol30-104:
To identify mutations causing synthetic lethality with pol30-104, we used a red/white colony sectoring assay (see MATERIALS AND METHODS). Of the 200,000 colonies screened, ten exhibited a Slp^- (synthetic lethality with pol30-104) phenotype. Backcrosses of the slp mutants with the unmutagenized parent strain revealed that each mutant carried a single recessive mutation causing the Slp^- phenotype. The ten slp mutations fell into five complementation groups. In addition to the dependency on POL30, mutations in groups 2, 3 and 4 caused sensitivities to MMS and HU.

The identities of complementation groups 1, 2, 3, and 4 were elucidated using a variety of methods. For group 1, the slp mutation was very closely linked to HIS3 (14PD:0NPD:1T), as is the gene for the large subunit of replication factor C (RFC1). Since RFC is responsible for loading PCNA onto DNA (FIEN and STILLMAN 1992 ; PODUST et al. 1992 ; TSURIMOTO and STILLMAN 1991 ), it seemed likely that group 1 mutations might affect the RFC1 gene. We determined that group 1 consists of rfc1 mutations by transforming strains CH2278 (slp1-1) and CH2280 (slp1-2) with plasmid pCH1545 (RFC1) and observing a relief of POL30 dependency. The slp mutations in groups 2 and 3 were identified by isolating genomic library plasmids that suppressed the MMS and HU sensitivities in strains CH2286 (slp2-1) and CH2297 (slp3). All library plasmids suppressing MMS and HU sensitivities also suppressed the POL30 dependency. Subcloning of the library inserts revealed that group 2 consists of rad50 mutations, and group 3 consists of a rad57 mutation. Finally, the group 4 mutation was identified as being in RAD52 by the failure of strain CH1637 (rad52) to complement the HU and MMS sensitivities of strain CH2294 (slp4). Thus, two classes of mutations, those in RFC1 and those in members of the RAD52 epistasis group, were identified as displaying synthetic lethality with pol30-104. While it is not surprising that rfc1 mutants were isolated in this screen, it is striking that mutations in three different members of the RAD52 epistasis group exhibit synthetic lethality with pol30-104.

Specificity of synthetic lethality:
To determine if pol30-104 displays synthetic lethality only with specific types of mutations affecting genes in the RAD52 epistasis group, we determined whether pol30-104 exhibits synthetic lethality with null alleles of members of the RAD52 epistasis group. Tetrad analysis was performed on diploids heterozygous for the pol30-104 allele and a rad51, rad52, or rad54 mutation. The pol30-104 mutation displayed synthetic lethality with rad52 (0 of 13 expected double mutants were viable) and rad54 (0 of 10 expected double mutants were viable). Six of the nine expected pol30-104 rad51 double mutants grew poorly at 35° and were inviable at 25°; the remaining three pol30-104 rad51 double mutants failed to grow even at 35°. Taken together, these crosses indicate that general disruptions in the RAD52 pathway are fatal to pol30-104 mutants.

To determine whether the synthetic lethality between pol30-104 and mutations in repair genes was specific to the RAD52 repair pathway, the viabilities of pol30-104 strains carrying mutations in other DNA repair genes were examined. Tetrad analysis was performed on diploids heterozygous for the pol30-104 allele and a deletion allele of rad1, rad14 (RAD3 epistasis group), or rad6 (RAD6 epistasis group). The pol30-104 mutation did not display synthetic lethality with rad1 (18 of 19 expected double mutants were viable), rad14 (7 of 7 expected double mutants were viable) nor rad6 (22 of 23 expected double mutants were viable). This observation rules out the possibility that pol30-104 strains require all DNA repair processes to be intact for viability, and it points to a specific requirement for the RAD52 pathway in pol30-104 strains.

To determine if the RAD52 pathway is required when cells carry general defects in DNA synthesis rather than defects specific to pol30-104 mutants, the viability of DNA synthesis mutants with a rad52 background was examined. Tetrad analysis was performed on diploids heterozygous for the rad52::URA3 mutation and a mutation in one of various DNA synthesis genes. The rad52 mutation did not display synthetic lethality with cdc17-1 (DNA polymerase ; 5 of 5 expected double mutants were viable), pol1-16 (DNA polymerase ; 8 of 8 expected double mutants were viable), pol2-11 (DNA polymerase ; 10 of 12 expected double mutants were viable), pol2-12 (DNA polymerase ; 16 of 18 expected double mutants were viable) nor cdc2-1 (DNA polymerase ; 15 of 22 expected double mutants were viable). These results are consistent with data demonstrating rad52 does not exhibit synthetic lethality with these mutations affecting the three essential yeast DNA polymerases (HARTWELL and SMITH 1985 ). In contrast, we found that rad52 does display synthetic lethality with rfc1-5 (0 of 46 expected double mutants were viable). This result is not surprising, because RFC is responsible for loading PCNA onto DNA (FIEN and STILLMAN 1992 ; LEE and HURWITZ 1990 ; LEE et al. 1988 ; PODUST et al. 1995 ; TSURIMOTO and STILLMAN 1991 ). Thus, it supports the independent observation of synthetic lethality between pol30-104 and mutations in genes belonging to the RAD52 epistasis group. Taken together, these results suggest that defects in the PCNA/RFC complex result in a requirement for the RAD52 pathway.

pol30 mutations exhibit allele-specific synthetic lethality with rad50:
While the previous experiments demonstrate that cells require an intact RAD52 pathway when there are defects in PCNA and RFC, other experiments are necessary to determine the specific defect in pol30-104 mutants that causes a requirement for an intact RAD52 pathway. For example, it is possible that any defect in Pol30p may cause a requirement for the RAD52 pathway. If so, one would expect that all pol30 mutations would exhibit synthetic lethality with rad50. Alternatively, the requirement for an intact RAD52 pathway may occur when a mutation in POL30 affects one particular in vivo function. If so, one would expect that pol30 mutations would exhibit an allele-specific synthetic lethality with rad50. Thus, determining the requirement for the RAD52 pathway in different pol30 mutants may prove useful in determining the defect in pol30-104 mutants that causes the requirement for the RAD52 pathway.

To determine whether the requirement for the RAD52 pathway is specific to certain pol30 alleles, we examined the viability of strains carrying various pol30 alleles in combination with rad50. To simplify the analysis, each strain initially carried a POL30 URA3 plasmid to ensure viability. The strains had the following genotype: pol30::LEU2 rad50::hisG ura3 trp1[pCH1511 (POL30 URA3)] [pCH1594 to pCH1608 (pol30-x TRP1)] (see Table 2). Serial dilutions of each strain were transferred onto 5-FOA plates to select for loss of the POL30 URA3 plasmid pCH1511. These plates revealed that synthetic lethality with rad50 occurs only with specific pol30 alleles (Figure 1). In group I, pol30-102, -105, -113 and -114 exhibited good viability in a rad50 background at 35° (25° and 14° data not shown); the pol30-105 mutation was unusual, in that it caused synthetic lethality at 14° (data not shown). In group II, pol30-100, -103, -104, -106, -108 and -112 displayed synthetic lethality with rad50 at 35°, 25° and 14°. It is most economical to conclude that these two groups of pol30 alleles differ in their detrimental effects on the cells. While defects caused by group I mutations appear to cause little need for Rad50p, defects caused by group II mutations cause a requirement for Rad50p. By identifying defects caused by each allele, we hoped to begin to understand what defect in Pol30p causes the requirement for an intact RAD52 pathway. Therefore, we next characterized pol30 mutant strains for a number of different phenotypes.

View larger version (71K): In this window In a new window Download PPT slide   Figure 1. Allele specificity of pol30 synthetic lethality with rad50. Strains of the general genotype pol30::LEU2 rad50::hisG ura3 trp1 [pCH1511 (POL30 URA3)] [pCH1594 to pCH1608 (pol30-x TRP1)] (see Table 2) were spotted onto synthetic complete (SC) and 5-FOA plates at 35°. 5-FOA selects for the loss of the POL30 URA3 plasmid. Group I pol30 alleles can support growth in the rad50 mutant strain, whereas group II pol30 mutations cause inviability in the rad50 background. With the exception of the inviability of pol30-105 rad50 at 14°, similar results were seen at 25° and 14° (data not shown).

Location of mutations in Pol30p:
To determine if a definable region of Pol30p is essential for viability in a rad50 background, the location of each pol30 mutation was mapped onto the three-dimensional structure of Pol30p (KRISHNA et al. 1994 ) (Figure 2). Group I mutations, which are viable in a rad50 background, are distributed among different regions of Pol30p, such as the inner surface of the protein (pol30-114), the interdomain region (pol30-102), and the ends of the monomers (pol30-113 and pol30-105). In contrast to the varied locations of group I mutations, all six of the group II pol30 mutations, which are inviable in a rad50 background, are located in the beta sheet interdomain region that bridges two homologous domains of each Pol30p monomer subunit. However, the pol30-102 mutation alters this same region and does not display synthetic lethality with rad50. Furthermore, cold-sensitive synthetic lethality is caused by the group I pol30-105 mutation that affects the monomer-monomer interface region of Pol30p. Thus, it appears unlikely that it is simply the location of a given mutation that determines viability in a rad50 background.

View larger version (33K): In this window In a new window Download PPT slide   Figure 2. Location of mutations on the crystal structure of Pol30p. The carbon backbone of the crystal structure of Pol30p monomer subunit is shown (KRISHNA et al. 1994 ). Arrows indicate the positions of amino acids affected by mutations. Group I mutations affect the inner surface (pol30-114), interdomain (pol30-102) and monomer-monomer interface (pol30-113 and pol30-105) regions of Pol30p. All six group II mutations affect amino acids in the interdomain region of Pol30p.

DNA repair capacity of pol30 mutants:
To determine whether synthetic lethality is the result of an additive defect in DNA repair caused by pol30 and rad50 mutations, the relative sensitivities to MMS and UV were determined for each pol30 mutant. None of the ten pol30 mutations in this study caused more than a slight sensitivity to UV (data not shown), and the mutations that exhibited a slight sensitivity to UV were distributed in group I and group II. In contrast, all pol30 mutations caused at least a moderate MMS-sensitive phenotype (Figure 3). Three of the group I mutations (pol30-102, -113, and -114) caused an intermediate sensitivity to MMS, and the pol30-105 mutation caused little sensitivity to 0.005% and 0.01% MMS. The group II alleles caused a wide range of sensitivities to MMS. Some group II mutations (pol30-100, -104, and -112) caused greater MMS sensitivity than that caused by group I mutations. Importantly, others (pol30-103, -106, and -108) caused a lesser sensitivity than the sensitivity caused by group I alleles. Since the degree of sensitivity to UV or MMS does not correlate with the observation of synthetic lethality with rad50, it is unlikely that it is simply an overall repair defect caused by pol30 mutations that causes a requirement for an intact RAD52 pathway.

View larger version (70K): In this window In a new window Download PPT slide   Figure 3. Methyl methane sulfonate sensitivity of pol30 mutants. Serial dilutions of pol30-x strains (strains CH2255 to CH2269; see Table 1) were spotted onto YEPD plates containing 0, 0.005 or 0.01% MMS. Plates were photographed after two days at 35°. Group I and group II mutations cause a wide range of sensitivity to MMS. Some group II mutations (pol30-100, -104, and -112) are more sensitive to MMS than group I mutations. Other group II mutations (pol30-103 and pol30-108) are less sensitive to MMS. Similar results were observed at 25° (data not shown).

Synthetic lethality with rad50 correlates with cell cycle alterations in pol30 mutants:
Microfluorometric analysis (FACS) was used to determine whether the pol30 mutations affect the overall ability to efficiently replicate the genome and proceed through the cell division cycle. Cultures growing exponentially at 35° were shifted to 25° or 14° for three generation times and then prepared for FACS analysis. The group I mutants exhibited FACS profiles that were indistinguishable from the POL30 strain at 35° (Figure 4). In contrast, all six group II pol30 mutations caused an accumulation of G2 cells, which varied in severity with temperature: two mutations (pol30-100 and pol30-112) caused an accumulation of G2 cells at all temperatures; four mutations (pol30-103, -104, -106, and -108) caused mild perturbations at 35° but profound alterations at 14°. In addition, the group I pol30-105 mutation yielded a normal FACS profile at 35° (the temperature at which it retains viability in a rad50 background), and it caused an enrichment of S and G2 cells at 14° (the temperature at which it exhibits synthetic lethality with rad50). These results reveal a correlation between the requirement for the RAD52 pathway in pol30 mutants and DNA replication defects caused by pol30 mutations.

View larger version (33K): In this window In a new window Download PPT slide   Figure 4. Microfluorometric (FACS) analysis of pol30 mutants. Cultures of pol30-x yeast strains were grown to log phase at 35°. Cultures were shifted to 25° or 14° for three generation times, and aliquots of each culture were fixed in ethanol and prepared for FACS. The FACS profile for each pol30 mutant is presented as a black outline superimposed on the FACS profile of the POL30 strain (gray-shaded silhouette) for each temperature. While group I mutations do not cause a detectable disruption of the cell division cycle, the group II mutations caused an accumulation of S and/or G2 cells. The single group I mutation, pol30-105, that caused a cold-sensitive synthetic lethality with rad50 caused a slight disruption in the cell division cycle when the temperature was decreased.

Synthetic lethality with rad50 correlates with an accumulation of small single stranded DNA fragments during DNA synthesis in vivo:
Since Pol30p is important for many DNA replication processes, the requirement for the RAD52 pathway in pol30 mutants could be caused by any of a number of DNA replication defects. However, it is striking that mutations affecting proteins required for Okazaki fragment maturation in vitro (DNA ligase, PCNA, RFC, and FEN-1) all display synthetic lethality with mutations affecting the RAD52 pathway (ISHIMI et al. 1988 ; JOHNSTON 1983 ; JOHNSTON and NASMYTH 1978 ; MONTELONE et al. 1981 ; TURCHI et al. 1994 ). An economical explanation for these synthetic lethal interactions is that the RAD52 pathway is required for completion of DNA replication when there are defects in Okazaki fragment maturation. This hypothesis predicts that group II pol30 mutations cause defects in Okazaki fragment maturation.

To determine if mutations affecting these DNA replication proteins affect the maturation of Okazaki fragments in vivo, alkaline sucrose velocity sedimentation was used to determine if mutants accumulate small ssDNA fragments during DNA synthesis in vivo. Exponentially growing cultures were labeled with ¹⁴C-uracil for at least six generation times to label the bulk of genomic DNA. A short pulse of ³H-uracil was then added to the cultures to differentially label newly synthesized DNA. The DNA was then isolated from these labeled cells and resolved by size by sedimentation through an alkaline sucrose density gradient. At 14°, group I mutants (pol30-102 and pol30-114) did not accumulate small ssDNA fragments and had sedimentation profiles similar to the POL30 strain (Figure 5 and Table 3). In contrast, group II pol30 mutants (pol30-100 and pol30-104) had a three- to fourfold increase in the accumulation of small ssDNA fragments during DNA synthesis at restrictive temperature. Furthermore, both rth1 and rfc1-1 mutations, which display synthetic lethality with rad52, also caused an accumulation of small ssDNA fragments at restrictive temperature. A concomitant decrease in the synthesis of large DNA molecules was observed at 14° in group II pol30 mutants and the rfc1-1 mutant. For the samples exhibiting an accumulation of small ssDNA, the DNA fragments in the ³H peak fractions corresponded in size to unligated Okazaki fragments (approximately 500 bp) (JOHNSTON and NASMYTH 1978 ; JOHNSTON and WILLIAMSON 1978 ). Thus, these results are consistent with the hypothesis that these mutations cause defects in the maturation of Okazaki fragments at restrictive temperature.

View larger version (25K): In this window In a new window Download PPT slide   Figure 5. Alkaline sucrose velocity sedimentation analysis of mutants pulse labeled at restrictive temperatures. To chronically label genomic DNA, yeast cultures were grown overnight in YEPD containing 0.08 µCi/ml 14C-uracil (dashed gray line). Cultures were shifted to restrictive temperature, and newly synthesized DNA was labeled with a pulse of 200 µCi/ml 3H-uracil (solid black line). Single-stranded DNA was resolved by size by sedimentation through an alkaline sucrose gradient as described in MATERIALS AND METHODS. The percent of total acid precipitable 14C (dashed gray line) and 3H (solid black line) CPM was determined for each of the 20 gradient fractions. Fraction #1 is the top of each gradient (small ssDNA fragments), and fraction #20 is the bottom of each gradient (large ssDNA fragments). A representative sedimentation profile is shown for each of the indicated strains. Top row: Sedimentation profiles for strains carrying mutations exhibiting synthetic lethality with rad50. Strains CH2425 (pol30-100), CH2427 (pol30-104), CH2429 (rth1), and CH2212 (rfc1-1) all accumulate short pieces of ssDNA during DNA synthesis in vivo. The fractions containing the peak 3H counts for pol30-100, pol30-104, rth1 and rfc1-1 mutants correspond to ssDNA that is ~500 bp in length. Bottom row: Sedimentation profiles are shown for group I pol30 mutants and the wild-type POL30 strain. Strains CH2426 (pol30-102) and CH2428 (pol30-114) mutants had sedimentation profiles similar to the wild-type CH2424 (POL30) profile.

If the requirement for the RAD52 pathway in Group II pol30 mutants is caused by defects in Okazaki fragment processing, group II pol30 mutations should cause an accumulation of small ssDNA molecules during DNA synthesis even at their "permissive" temperature (a temperature at which they display synthetic lethality with rad50). To test this hypothesis, DNA was isolated from pol30 mutants that were pulse-labeled at 35°, and it was analyzed on alkaline sucrose velocity sedimentation gradients (Figure 6). The profiles for Group I pol30 mutants (pol30-102 and pol30-114) were similar to the profiles for POL30 strain. In contrast, group II mutations (pol30-100 and pol30-104) caused an accumulation of small ssDNA fragments in newly synthesized DNA (19% and 22%, respectively) (Table 3). In addition, the amount of ³H label in the bottom nine fractions (56% for pol30-100 and 57% for pol30-104) indicate that these mutants are capable of synthesizing large DNA molecules at 35°. While the defects observed at 35° were substantially less than those observed at 14°, the accumulation of small ssDNA fragments suggests that these mutations cause defects in lagging-strand DNA synthesis even at "permissive" temperatures. Thus, these results are consistent with the hypothesis that the requirement for an intact RAD52 pathway is caused by defects in Okazaki fragment maturation.

View larger version (25K): In this window In a new window Download PPT slide   Figure 6. Alkaline sucrose velocity sedimentation analysis of pol30 mutants at 35°. Cultures were treated as described in Figure 5, except that pulse labeling of newly synthesized DNA with 3H-uracil was performed at "permissive" temperature (35°) for 30 min. (see MATERIALS AND METHODS). Group I pol30 strains (pol30-102 and pol30-114) had sedimentation profiles that were very similar to the wild-type POL30 strain at 35°. In contrast, both group II mutations (pol30-100 and pol30-104) caused an accumulation of small ssDNA fragments during DNA synthesis in vivo even at the "permissive" temperature.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Using a colony-sectoring assay, we identified mutations affecting members of the RAD52 epistasis group that display synthetic lethality with pol30-104. This synthetic lethality is specific, because pol30-104 mutants do not require the function of proteins belonging to the RAD3 or RAD6 epistasis groups. The RAD52 pathway does not appear to be required when there are defects in general DNA replication and repair processes caused by mutations in any of the three essential DNA polymerases (Pol, Pol and Pol). The allele specificity of synthetic lethality demonstrates that it is linked to a disruption in the cell division cycle caused by pol30 mutations. Furthermore, pol30 alleles that exhibit synthetic lethality with rad50 also exhibit an accumulation of short single strands in newly synthesized DNA.

The outcome of the screen for synthetic lethality suggests that we achieved good recovery of candidate mutations. Three of the nine identified mutations affect the structural gene for replication factor C. The RFC complex and Pol30p physically interact (FIEN and STILLMAN 1992 ; PODUST et al. 1995 ; TSURIMOTO and STILLMAN 1991 ), and mutations in rfc1 and pol30 have previously been shown to display synthetic lethality with each other (AMIN and HOLM 1996 ). Thus, the recovery of rfc1 mutations shows that the screen was effective in identifying known proteins that functionally interact with Pol30p. Although all of the other identified mutations affect members of the RAD52 epistasis group, we did not recover mutations in every member of the RAD52 epistasis group in the screen. Since additional crosses revealed that rad51 and rad54 deletions each exhibit synthetic lethality with pol30-104, it is clear that the screen was not completely saturated. Even though this screen did not generate a comprehensive collection of mutations displaying synthetic lethality with pol30, it is striking that six of the nine identified slp mutations were in members of the RAD52 epistasis group. This clustering of mutations to one functional pathway indicates that cells with partially defective Pol30p have a low tolerance for defects in recombinational repair.

The DNA sedimentation profiles in this study are consistent with previously published sedimentation profiles. Previously, only chronically labeled DNA was examined in sedimentation velocity gradients; only small differences were observed between mutant and wild-type (<25% change in the number-average molecular weight of DNA), even when pol30 and rfc1 mutants were grown at their restrictive temperature (AMIN and HOLM 1996 ; MCALEAR et al. 1996 ; and M. TUFFO and C. HOLM, unpublished data). The chronically labeled DNA in the present study shows a similar pattern to the previously published profiles (dashed lines in Figure 5 and Figure 6). In contrast, when newly synthesized DNA is labeled with a pulse of ³H-uracil, these same mutations cause substantial effects on the accumulation of small ssDNA fragments (a two- to fourfold increase in small ssDNA). This striking difference between the results with chronically labeled DNA versus pulse-labeled DNA suggests that the normal processing of small DNA fragments into large DNA strands is defective in these mutants.

Taken together with other published data, the results of this synthetic lethal screen provide additional insight into the role of Pol30p in mutation avoidance. Epistasis analysis of the spontaneous mutation rates caused by pol30 mutations, rth1 mutations, and mutations in members of the mismatch repair (MMR) pathway suggested that Pol30p and Rth1p may play a role in the MMR pathway (JOHNSON et al. 1996 ; JOHNSON et al. 1995 ). However, a detailed analysis revealed that rth1 mutations cause fundamentally different types of spontaneous mutations than those caused by inactivation of the MMR pathway (TISHKOFF et al. 1997 ). One interpretation is that Rth1p may not play a major role in the MMR pathway, and it may be the initial member of a previously undefined mutation avoidance pathway. It is unclear whether Pol30p may also be a member of this new pathway. Interestingly, the RAD52 pathway appears to be essential for viability when this new pathway is inactivated, because rth1 displays synthetic lethality with rad51 and rad52 mutations (TISHKOFF et al. 1997 ); in contrast, the RAD52 pathway is not required in yeast strains carrying mutations affecting the MMR pathway (SAPARBAEV et al. 1996 ). Thus, our findings that pol30 and rfc1 mutations also display synthetic lethality with rad52 suggest that RFC and PCNA could be involved in the RTH1 mutation avoidance pathway.

In addition to pol30, rfc1, and rth1, mutations affecting Cdc9p and Rad3p also display synthetic lethality with rad52 (MALONE and HOEKSTRA 1984 ; MONTELONE et al. 1981 ). The mechanism by which rad3 mutations display synthetic lethality with rad52 is probably different from that of mutations affecting the other four genes. Synthetic lethality is only observed between rad52 and a specific class of rad3 mutations. These Rem^- rad3 mutations cause hyper-recombination and mutator phenotypes (MALONE and HOEKSTRA 1984 ; MONTELONE et al. 1988 ). In contrast to the recessive pol30, rfc1, rth1, and cdc9 mutations that display synthetic lethality with rad52, the Rem^- rad3 mutations are semi-dominant and cause gain of function phenotypes (MALONE and HOEKSTRA 1984 ; MONTELONE et al. 1988 ). Furthermore, the inviability of rad3 rad52 mutants is suppressed by mutational inactivation of Rad1p or Rad4p endonucleases (MONTELONE et al. 1988 ). Thus, it is thought that Rem^- rad3 mutations cause an activation of Rad1p and Rad4p endonucleases, and the activated endonucleases create DNA damage that is lethal to rad52 mutants.

Interestingly, the other four gene products (Pol30p, Rfc1p, Rth1p, and Cdc9p) known to be required in rad52 mutants all are important for lagging-strand DNA synthesis in vitro. Rth1p has 5' flap endonuclease and 5'-3' exonuclease activities that may be particularly important for processing Okazaki fragments (ISHIMI et al. 1988 ; LI et al. 1995 ; TURCHI et al. 1994 ; WU et al. 1996 ). Pol30p physically interacts with Rth1p, and at high concentrations, Pol30p can stimulate its endonuclease activity on model flap structures and its exonuclease activity on nicked DNA in vitro (LI et al. 1995 ; WU et al. 1996 ). This latter stimulation is dependent upon replication factor C (LI et al. 1995 ). These observations suggest that the maturation of Okazaki fragments in vivo could be disrupted by mutations affecting either Pol30p itself or RFC-mediated loading of Pol30p onto DNA. Finally, the DNA ligase activity of Cdc9p is critical for ligating ssDNA nicks during the maturation of Okazaki fragments ( JOHNSTON 1983 ; JOHNSTON and NASMYTH 1978 ).

Our observations that mutations affecting proteins required for the in vitro reconstitution of lagging-strand DNA synthesis (pol30-100, pol30-104, rfc1-1, and rth1) all cause an accumulation of small ssDNA fragments during DNA synthesis in vivo strongly suggest that Okazaki fragment maturation is defective in these mutants in vivo. Since small ssDNA fragments accumulate only in pulse-labeled DNA and not in chronically labeled DNA, these ssDNA fragments are the result of defects in newly synthesized DNA. Furthermore, the sedimentation profile of pulse-labeled DNA from a cdc9 mutant is very similar to what we observed with pol30-100, pol30-104, rfc1-1 and rth1 mutants (JOHNSTON and NASMYTH 1978 ). Therefore, the most economical explanation is that the small ssDNA fragments are unligated (or partially ligated) lagging-strand DNA synthesis intermediates, and that these mutations affect Okazaki fragment processing.

Although mutations displaying synthetic lethality with rad52 may cause defects in leading-strand synthesis, it is unlikely that these defects are the cause of the synthetic lethality. Less than 50% of ³H counts were recovered from the bottom nine gradient fractions (DNA greater than 50 kb) of group II pol30 mutants at 14°. This observation suggests that these mutants also have defects in leading strand DNA synthesis at their restrictive temperature, and it is consistent with the observation that the pol30-104 mutant progresses slowly through S-phase at 14° (AMIN and HOLM 1996 ). In contrast, even though the group II pol30-104 mutant accumulates small ssDNA fragments at 35°, it is capable of synthesizing large DNA molecules at 35° (57% of ³H counts in the bottom nine gradient fractions), and it progresses through S-phase with normal kinetics at this temperature (AMIN and HOLM 1996 ). Thus, it appears that the RAD52 pathway is required when there are defects in lagging-strand DNA synthesis.

The precise role of the RAD52 pathway in mutants with defects in DNA replication remains unclear, but data presented here suggest some interesting possibilities. rfc1-1, rth1, and group II pol30 mutations all cause an accumulation of single-stranded nicks or gaps, which are probably the result of defective Okazaki fragment maturation. These defects cause a requirement for the RAD52 recombinational repair pathway. One possible explanation for this requirement is that the RAD52 pathway may compensate for these defects by compensating for inefficient Okazaki fragment processing through a post-replicative repair mechanism. Replication of DNA that is normally synthesized by lagging-strand DNA synthesis could be initiated at the 3' end of a newly synthesized leading strand that has undergone a strand invasion into its sister chromatid. Elongation of this 3' end could be accomplished by using the newly synthesized leading strand of the sister chromatid as a template. This mechanism would allow regions of a genome normally replicated by lagging-strand (discontinuous) synthesis to be replicated in a continuous manner. Alternatively, the RAD52 pathway may play a less direct role in compensating for defects in Okazaki fragment maturation. The abundance of single-stranded breaks in the lagging strand of cdc9, rfc1, rth1 and pol30 mutants may present a particularly dangerous situation for rad52 mutants. Any lesion in the template for the lagging strand would result in a double-stranded break that would be lethal to a rad52 mutant.

While the interaction between DNA replication and recombination processes is not a novel concept, the mechanisms of the interaction remain elusive. For over twenty years, it has been known that recA polA double mutants in E. coli are inviable (GROSS et al. 1971 ; MONK and KINROSS 1972 ), and recently it has been discovered that both the 5' exonuclease and DNA polymerase activities of PolA are required in recombination-deficient recA mutants (CAO and KOGAMA 1995). Furthermore, extensive DNA synthesis is observed during recombinational repair in both yeast and bacteria (ASAI et al. 1994 ; MALKOVA et al. 1996 ). While recombinational repair proteins are not required for the in vitro reconstitution of DNA replication, Rad50p and Rad52p have been suggested to play a role during S phase because both are required for replication slippage caused by the pol3-t mutation (TRAN et al. 1995 ), and both are required in response to the DNA synthesis inhibitor, hydroxyurea. Data presented here suggest that rfc1, rth1, and certain pol30 mutations cause defects in Okazaki fragment maturation, and that the RAD52 pathway must be intact to compensate for these defects. Thus, it appears that the RAD52 pathway may act as a fail safe mechanism that enables proliferating cells to recover from potentially lethal defects in the process of DNA replication.

	ACKNOWLEDGMENTS

We thank MIKE WILSON for his assistance with the synthetic lethal screen; NEELAM AMIN and members of C. HOLM's lab for their suggestions relating to this work; DAVID SCHILD and SATYA PRAKASH for plasmids; and SATYA PRAKASH and NEELAM AMIN for their critical reading of the manuscript. This work was supported by grant GM-36510 from the National Institutes of Health (NIH). B.J.M. was partially supported by an NIH training grant CA-67754.

Manuscript received April 8, 1997; Accepted for publication October 20, 1997.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AMIN, N. S. and C. H. HOLM, 1996  In vivo analysis reveals that the interdomain region of the yeast proliferating cell nuclear antigen is important for DNA replication and DNA repair. Genetics 144:479-493[Abstract].

ASAI, T., D. B. BATES, and T. KOGOMA, 1994  DNA replication triggered by double-stranded breaks in E. coli: dependence on homologous recombination functions. Cell 78:1051-1061[Medline].

AUSUBEL, F. M., R. BRENT, R. E. KINGSTON, D. D. MOORE, J. G. SEIDMAN et al. (Editors), 1988 Current Protocols in Molecular Biology. Greene Publishing Associates and Wiley-Interscience, NY.

AYYAGARI, R., K. J. IMPELLIZZERI, B. L. YODER, S. L. GARY, and P. M. J. BURGERS, 1995  A mutational analysis of the yeast proliferating cell nuclear antigen indicates distinct roles in DNA replication and DNA repair. Mol. Cell. Biol. 15:4420-4429[Abstract].

BUDD, M. E. and J. L. CAMPBELL, 1993  DNA polymerases and are required for chromosomal replication in Saccharomyces cerevisiae.. Mol. Cell. Biol. 13:496-505[Abstract/Free Full Text].

BUDD, M. E., K. D. WITTRUP, J. E. BAILEY, and J. L. CAMPBELL, 1989  DNA polymerase I is required for promeiotic DNA replication and sporulaton but not for X-ray repair in Saccharomyces cerevisiae.. Mol. Cell. Biol. 9:365-376[Abstract/Free Full Text].

BURGERS, P. M., 1991  Saccharomyces cerevisiae replication factor C: II. Formation and activity of complexes with the proliferating cell nuclear antigen and with DNA polymerases and . J. Biol. Chem. 266:22698-22706[Abstract/Free Full Text].

BURGERS, P. M. and B. L. YODER, 1993  ATP-independent loading of the proliferating cell nuclear antigen requires DNA ends. J. Biol. Chem. 268:19923-19926[Abstract/Free Full Text].

CAO, Y. and T. KOGOMA, 1995  The mechanism of recA polA lethality: suppression by RecA-independent recombination repair activated by the lexA(Def ) mutation in Escherichia coli.. Genetics 139:1483-1494[Abstract].

FIEN, K. and B. STILLMAN, 1992  Identification of replication factor C from Saccharomyces cerevisiae : a component of the leading-strand DNA replication complex. Mol. Cell. Biol. 12:155-163[Abstract/Free Full Text].

GROSS, J. D., J. GRUNSTEIN, and E. M. WITKIN, 1971  Inviability of recA^- derivatives of the DNA polymerase mutant of De Lucia and Cairns. J. Mol. Biol. 58:631-634[Medline].

HARTWELL, L. H., R. K. MORTIMER, J. CULOTTI, and M. CULOTTI, 1973  Genetic control of the cell division cycle in yeast: V. Genetic analysis of cdc mutants. Genetics 74:267-286[Abstract/Free Full Text].

HARTWELL, L. H. and D. SMITH, 1985  Altered fidelity of mitotic chromosome transmission in cell cycle mutants of S. cerevisiae.. Genetics 110:381-395[Abstract/Free Full Text].

ISHIMI, Y., A. CLAUDE, P. BULLOCK, and J. HURWITZ, 1988  Complete enzymatic synthesis of DNA containing the SV40 origin of replication. J. Biol. Chem. 263:19723-19733[Abstract/Free Full Text].

JOHNSON, R. E., G. K. KOVVALI, S. N. GUZDER, N. S. AMIN, and C. HOLM et al., 1996  Evidence for involvement of yeast proliferating cell nuclear antigen in DNA mismatch repair. J. Biol. Chem. 271:27987-27990[Abstract/Free Full Text].

JOHNSON, R. E., G. K . KOVVALI, L. PRAKASH, and S. PRAKASH, 1995  Re-quirement of the yeast RTH1 5' to 3' exonuclease for the stability of simple repetitive DNA. Science 269:238-240[Abstract/Free Full Text].

JOHNSTON, L. H., 1983  The cdc9 ligase joins completed replicons in baker's yeast. Mol. Gen. Genet. 190:315-317[Medline].

JOHNSTON, L. H. and K . A. NASMYTH, 1978  Saccharomyces cerevisiae cell cycle mutant cdc9 is defective in DNA ligase. Nature 274:891-893[Medline].

JOHNSTON, L. H. and D. H. WILLIAMSON, 1978  An alkaline sucrose gradient analysis of the mechanism of nuclear DNA synthesis in the yeast Saccharomyces cerevisiae.. Mol. Gen. Genet. 164:217-225[Medline].

KRANZ, J. E. and C. HOLM, 1990  Cloning by function: an alternative approach for identifying yeast homologs of genes from other organsms. Proc. Natl. Acad. Sci. USA 87:6629-6633[Abstract/Free Full Text].

KRISHNA, T. S. R., X. P. KONG, S. GARY, P. BURGERS, and J. KURIYAN, 1994  Crystal structure of the eukaryotic DNA polymerase processivity factor PCNA. Cell 79:1233-1243[Medline].

LEE, S.-H. and J. HURWITZ, 1990  Mechanism of elongation of primed DNA by DNA polymerase delta, proliferating cell nuclear antigen and activator I. Proc. Natl. Acad. Sci. USA 87:5672-5676[Abstract/Free Full Text].

LEE, S.-H., Y. ISHIMI, M. K. KENNY, P. BULLOCK, and F. B. DEAN et al., 1988  An inhibitor of the in vitro elongation reaction of simian virus 40 DNA replication is overcome by proliferating-cell nuclear antigen. Proc. Natl. Acad. Sci. USA 85:9469-9473[Abstract/Free Full Text].

LEE, S.-H., Z.-Q. PAN, A. D. KWONG, P. M. J. BURGERS, and J. HURWITZ, 1991  Synthesis of DNA by DNA polymerase in vitro.. J. Biol. Chem. 266:22707-22717[Abstract/Free Full Text].

LEVIN, D. and F. HUTCHINSON, 1973  Relation between single-strand DNA mass and sedimentation distance in alkaline sucrose gradients. J. Mol. Biol. 75:495-502[Medline].

LI, X., J. LI, J. HARRINGTON, M. R. LIEBER, and P. M. BURGERS, 1995  Lagging-strand DNA synthesis at the eukaryotic replication fork involves binding and stimulation of FEN-1 by proliferating cell nuclear antigen. J. Biol. Chem. 270:22109-22112[Abstract/Free Full Text].

MALKOVA, A., E. L. IVANOV, and J. E. HABER, 1996  Double-strand break repair in the absence of RAD51 in yeast: a possible role for break-induced DNA replication. Proc. Natl. Acad. Sci. USA 93:7131-7136[Abstract/Free Full Text].

MALONE, R. E. and M. F. HOEKSTRA, 1984  Relationships between a hyper-rec mutation (REM1) and other recombination and repair genes in yeast. Genetics 107:33-48[Abstract/Free Full Text].

MATSUMOTO, Y., K . KIM, and D. F. BOGENHAGEN, 1994  Proliferating cell nuclear antigen-dependent abasic site repair in Xenopus laevis oocytes: an alternative pathway of base excision DNA repair. Mol. Cell. Biol. 14:6187-6197[Abstract/Free Full Text].

MCALEAR, M., K . M. TUFFO, and C. HOLM, 1996  The large subunit of replication factor C (Rfc1/Cdc44p) is required for DNA replication and DNA repair in S. cerevisiae.. Genetics 142:65-78[Abstract].

MONK, M. and J. KINROSS, 1972  Conditional lethality of recA and recB derivatives of a strain of Escherichia coli K-12 with a temperature-sensitive deoxyribonucleic acid polymerase I. J. Bacteriol. 109:971-978[Abstract/Free Full Text].

MONTELONE, B. A., M. F. HOEKSTRA, and R. E. MALONE, 1988  Sponta-neous mitotic recombination in yeast: the hyper-recombinational rem1 mutations are alleles of the RAD3 gene. Genetics 119:289-301[Abstract/Free Full Text].

MONTELONE, B. A., S. PRAKASH, and L. PRAKASH, 1981  Spontaneous mitotic recombination in mms8-1, an allele of the CDC9 gene of Saccharomyces cerevisiae.. J. Bacteriol. 147:517-525[Abstract/